C H A P T E R 1
you know only English, you will not receive the message I am sending. If I am
speaking Japanese grammar with English vocabulary, your understanding will
also be seriously impeded. There has to be agreement about the syntax (the
combination of the words), the vocabulary (the mapping between symbol and
concept), and the semantics (the meaning of the vocabulary). RDF is a set of
rules for creating semantics, and RDF Schema is a way of creating vocabularies.
In practice, however, you rarely need to get that philosophical. You do have to
recognize that classifications are necessarily arbitrary and that the names of
concepts are inherently meaningless. Any knowledge representation scheme
will have to take this arbitrariness into account.
Representing knowledge in a computerized format has been high fashion in the
computer industry for a long time. Knowledge representation systems are an
important aspect of metadata, but they have a big inherent weakness. While
they are mostly geared toward the scientific domain, they assume they are gen-
eral, both in vocabulary and structure. Of course, they are not, and while they
have come up with a lot of valid results in the field of intelligent agents, the work
still suffers from the assumptions that there are universal ways of representing
knowledge within one, centrally determined framework. This is most evident in
the classification schemes used in the library world. Classifying, as they do,
information objects that represent anything, they do have presumptions to be
general. But they are really specific, not in the least culture-specific. A library in
Sweden is not classified in the same way as a library in the United States.
RDF addresses this by providing for a decentralized scheme. Using the same
structure, it allows anyone to create their own vocabulary. This enables you to
take the good things from the artificial intelligence community (intelligent
agents and reasoning systems) and apply the good things from the Web (the
transport system, the data representation, the decentralization and indepen-
dence from a particular system). This intersection is the topic for this book.
What Is RDF?
RDF is a format to make assertions statements that are intended to point
something out. It has two roots: metadata and knowledge representation. The
concepts do overlap, but the technologies which have been developed in the
two fields, and the understanding of them, are very different.
Library communities have struggled to come up with a universal format for
describing books that can be used in electronic catalogs. The result is the
Dublin Core format (named after Dublin, Ohio, and not Dublin, Ireland). It is
basically a set of attribute names and rules for which values they can take. The
69528_CH01Ix 4/6/2001 8:15 AM Page 4